Understanding the impact of city-wide cycling corridors on cycling mode share among different demographic clusters in Greater Melbourne, Australia

Abstract

Abstract In car-dominated cities like Melbourne, Australia, limited data on cyclists' travel patterns and socio-demographic differences complicate understanding the effectiveness of infrastructure investment interventions aimed at promoting cycling. Recent advancements in city-scale transport modelling enable virtual testing of such interventions. However, the application of agent- and activity-based models for large-scale cycling simulations has been constrained by data and complexity. In this study, we developed a city-scale agent-based simulation model for Greater Melbourne to evaluate changes in travel mode share from cycling infrastructure modifications. We clustered bicycle riders into five demographic groups: Maverick Males, Motivated Adults, Conscientious Commuters, Young Sprinters, and Relaxed Cruisers, estimating mode choice parameters for each group. Using aggregated smartphone application data, we developed a cycling trip routing methodology to incorporate road infrastructure impacts. Results indicated that travel time significantly influences mode choice across all clusters. Cycling infrastructure was crucial for four clusters, and travel cost influenced four clusters. The calibrated model assessed the potential impact of fully implementing Greater Melbourne's strategic cycling corridors, a network of key cycling routes. Simulations suggested an initial 30% increase in cycling use, raising the mode share to approximately 2.6%, indicating a modest overall impact. Further analysis showed that even with full implementation, on average about half of the lengths of the routed bikeable trips would still occur on roads without any cycling infrastructure. This underscores the need to improve infrastructure on both major corridors and minor roads, and to complement these improvements with behavioural interventions.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 3 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.